Will Chen
6c82cef5b5
add Qwen3 coder ( #808 )
2025-08-04 17:04:09 -07:00
Will Chen
d56006f847
Add kimi k2 ( #662 )
2025-07-17 21:20:32 -07:00
Will Chen
5d678c2ead
Smart auto ( #476 )
2025-06-23 23:08:29 -07:00
Will Chen
897d2e522c
enable engine for all models ( #434 )
2025-06-18 17:17:29 -07:00
Will Chen
e326f14779
Revert "Revert "Update gemini 2.5 models to GA variants ( #425 )" ( #429 )" ( #437 )
...
This reverts commit ff4e93d747
2025-06-18 17:11:17 -07:00
Will Chen
5a35c427ba
Add R1 ( #436 )
2025-06-18 12:34:52 -07:00
Will Chen
ff4e93d747
Revert "Update gemini 2.5 models to GA variants ( #425 )" ( #429 )
...
This reverts commit d69284e3b0
2025-06-17 12:57:43 -07:00
Will Chen
5694934183
Lower o3/o4 mini max output tokens to maximize input tokens ( #424 )
2025-06-17 10:31:24 -07:00
Will Chen
d69284e3b0
Update gemini 2.5 models to GA variants ( #425 )
...
Co-authored-by: graphite-app[bot] <96075541+graphite-app[bot]@users.noreply.github.com>
2025-06-17 10:24:54 -07:00
Will Chen
30b5c0d0ef
Replace thinking with native Gemini thinking summaries ( #400 )
...
This uses Gemini's native [thinking
summaries](https://cloud.google.com/vertex-ai/generative-ai/docs/thinking#thought-summaries )
which were recently added to the API.
Why? The grafted thinking would sometimes cause weird issues where the
model, especially Gemini 2.5 Flash, got confused and put dyad tags like
`<dyad-write>` inside the `<think>` tags.
This also improves the UX because you can see the native thoughts rather
than having the Gemini response load for a while without any feedback.
I tried adding Anthropic extended thinking, however it requires temp to
be set at 1, which isn't ideal for Dyad's use case where we need precise
syntax following.
2025-06-16 17:29:32 -07:00
Will Chen
ddce3c65d3
Lower max tokens for anthropic ( #367 )
2025-06-09 13:44:13 -07:00
Will Chen
9e50aa259e
Add Claude Sonnet 4 + o4-mini ( #237 )
2025-05-23 15:45:20 -07:00
Will Chen
0f6f069e43
Update Flash to 05-20 ( #226 )
2025-05-22 10:37:26 -07:00
Will Chen
2455c554ee
Add sonnet 3.5 ( #182 )
2025-05-16 11:06:43 -07:00
Will Chen
9a288fd82e
Fix local models ( #176 )
...
Fixes #167
2025-05-15 16:46:29 -07:00
Will Chen
09fc028f94
Mark which models are eligible for turbo edits ( #172 )
2025-05-15 16:02:42 -07:00
Will Chen
16e1dd7fc4
Add Claude Haiku 3.5 ( #171 )
2025-05-15 15:41:14 -07:00
Will Chen
d545babb63
Quick fix for Google models ( #160 )
2025-05-13 22:21:27 -07:00
Will Chen
ee5865dcf8
Precise custom model selection & simplify language model/provider log… ( #147 )
...
…ic (no merging)
2025-05-12 23:24:39 -07:00
Will Chen
d1027622b4
Loosen provider type to a string ( #144 )
2025-05-12 22:29:36 -07:00
Will Chen
f628c81f4c
Refactor constants/models and inline ( #143 )
2025-05-12 22:20:16 -07:00
Will Chen
877c8f7f4f
Simplify provider logic and migrate getContextWindow ( #142 )
2025-05-12 22:18:49 -07:00
Will Chen
993c5417e3
Fix DB schemas ( #138 )
2025-05-12 21:51:08 -07:00
Will Chen
e115074937
Fix DB schema ( #135 )
2025-05-12 16:12:36 -07:00
Will Chen
477015b43d
allow creating and listing custom language model ( #134 )
2025-05-12 16:00:16 -07:00
Will Chen
cd7eaa8ece
Prep for custom models: support reading custom providers ( #131 )
2025-05-12 14:52:48 -07:00
